KR101477051B1 - Method for object recognition through image-code transformation - Google Patents

Abstract

The present invention relates to a method for coding an image of an object, including the steps of: extracting feature points of an object; and generating an information code to define the object using the feature points. Also, the present invention relates to a method and a device for recognizing an object using information code to define an object.

Description

{METHOD FOR OBJECT RECOGNITION THROUGH IMAGE-CODE TRANSFORMATION}

The present invention relates to an image coding method for object recognition.

Recently, with the rapid spread of smart mobile devices and the practical use of wireless networks, there has been a rapid increase in demands for realistic services as well as static services. One of the technologies for realistic services is augmented reality. Augmented reality refers to a technique of superimposing a three-dimensional virtual image on a realistic image or background and displaying it as a single image.

Actual augmented reality is not well utilized in reality compared with the study on augmented reality. One of the main reasons why Augmented Reality is not well utilized is that it is not suitable for real-time service.

SURF and FAST algorithms are used for real-time service recognition. In smartphone environment, object recognition performance of both algorithms is only 1 ~ 4 fps (frame per sec). In order for users to experience real-time service, minimum 10 fps performance must be guaranteed, so existing algorithms are not suitable for providing real-time service.

Therefore, in order to provide a real-time sensible service in a mobile environment, a method of recognizing an object more quickly is needed.

It is an object of the present invention to provide a method of encoding an image of an object in order to realize object recognition in real time in a mobile environment.

The technical problem of the present invention is not limited to the above-mentioned technical problems, and other technical problems which are not mentioned can be clearly understood by those skilled in the art from the following description.

The image coding method of an object according to an embodiment of the present invention may include extracting minutiae points of an object and generating an information code defining the object using the minutiae points.

The method may further include selecting an optimal feature point among the extracted feature points after extracting the feature points of the object.

According to an embodiment of the present invention, the step of selecting an optimal feature point among the extracted feature points may include classifying the feature points that are concentrated among the extracted feature points, and classifying the feature points that are closest to the color average value among the classified feature points And selecting the feature point as the optimal feature point.

According to an embodiment, generating the information code may include calculating a distance between the minutiae points.

According to an embodiment, generating the information code may further comprise calculating a color average value around the feature points.

According to an embodiment of the present invention, the method may further include generating an information code of the object using a distance between the minutiae points and a color mean value around the minutiae points.

According to an embodiment of the present invention, the step of generating the information code may further include obtaining at least one of resolution of the image, context information of the object, and division information of the feature points.

According to the embodiment, at least two or more of the minutiae points of the extracted object, the distance between the minutiae points, the color average value around the minutiae points, the resolution of the image, the context information about the object, And generating an information code of the object.

An object recognition method according to an embodiment of the present invention includes extracting minutiae points of an object, extracting an information code defining the object using the minutiae points, and extracting an information code from the database storing the image using the extracted information code And searching for the object.

According to an embodiment of the present invention, the information code includes at least one of an optimal feature point of the object, a distance between the feature points, a color average value around the feature points, a resolution of the image, context information about the object, And may include two or more.

According to an embodiment of the present invention, the step of searching for an object from a database storing an image using the extracted information code may include a first filtering step using context information of the object, a second filtering step using division information of the feature points Comparing the optimal feature points, comparing distances between the feature points, and comparing color mean values around the feature points.

The recording medium according to the embodiment of the present invention may record a method of image coding of the object and a program for performing the object recognition method.

An information code of an object according to an embodiment of the present invention includes a feature point field for storing feature points of an object, a feature point distance field for storing a distance between feature points of the object, and a feature point color average value field for storing a color mean value around the feature points . ≪ / RTI >

According to an embodiment of the present invention, the apparatus may further include a resolution field for storing the resolution of the image, a context information field for storing context information about the object, and a partition information field for storing partition information of the object feature points.

An object recognition apparatus according to an embodiment of the present invention includes a feature point extraction unit that extracts feature points of the object from an image of an input object and an information code generation unit that generates an information code that defines the object from the extracted feature points .

According to the embodiment, the information code generating unit may include a distance calculating unit for calculating the distance between the extracted feature points, and a color average value calculating unit for calculating a color average value around the feature points.

According to an embodiment of the present invention, the information code generating unit includes a resolution calculating unit for calculating a resolution of an image of the object, a position information recognizing unit for obtaining position information of the object, and a feature point position analyzing unit for analyzing a position of the feature points in the image of the object. And may further include an analysis section.

According to an embodiment, the information processing apparatus may further include a database storing the generated information codes.

According to an embodiment of the present invention, the image recognition apparatus may further include an object recognition unit for searching the database for an image of the input object, wherein the object recognition unit compares the information code of the image of the input object with the information code stored in the database, have.

According to an aspect of the present invention, a time required for object recognition can be reduced by recognizing an object by encoding an image of the object.

According to an aspect of the present invention, an object can be recognized in real time in a mobile environment.

The effects of the present invention are not limited to the above-mentioned effects, and the effects not mentioned can be clearly understood by those skilled in the art from the present specification and the accompanying drawings.

FIG. 1 is a diagram for explaining an image coding method of an object according to an embodiment of the present invention.
2 shows a result of extracting a feature point from an image of an object.
FIG. 3 shows the result of selecting optimal feature points from the feature points shown in FIG. 2. FIG.
FIG. 4 shows the distances between the optimal feature points selected in FIG.
5 is a diagram for explaining a method of obtaining a color average value around a minutia.
6 is a diagram for explaining division information of minutiae points.
7 shows an information code generated by an image coding method according to an embodiment of the present invention.
8 is a view for explaining an object recognition method according to an embodiment of the present invention.
FIG. 9 illustrates an object recognition apparatus according to an embodiment of the present invention.

Other advantages and features of the present invention and methods of achieving them will become apparent with reference to the embodiments described below in detail with reference to the accompanying drawings. The present invention may, however, be embodied in many different forms and should not be construed as being limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the invention to those skilled in the art. Is provided to fully convey the scope of the invention to those skilled in the art, and the invention is only defined by the scope of the claims.

Unless defined otherwise, all terms (including technical or scientific terms) used herein have the same meaning as commonly accepted by the generic art in the prior art to which this invention belongs. Terms defined by generic dictionaries may be interpreted to have the same meaning as in the related art and / or in the text of this application, and may be conceptualized or overly formalized, even if not expressly defined herein I will not.

The terminology used herein is for the purpose of illustrating embodiments and is not intended to be limiting of the present invention. In the present specification, the singular form includes plural forms unless otherwise specified in the specification. As used herein, the terms' comprise 'and / or various forms of use of the verb include, for example,' including, '' including, '' including, '' including, Steps, operations, and / or elements do not preclude the presence or addition of one or more other compositions, components, components, steps, operations, and / or components.

The term 'and / or' as used herein refers to each of the listed configurations or various combinations thereof.

It should be noted that the terms such as '~', '~ period', '~ block', 'module', etc. used in the entire specification may mean a unit for processing at least one function or operation. For example, a hardware component, such as a software, FPGA, or ASIC. However, '~ part', '~ period', '~ block', '~ module' are not meant to be limited to software or hardware. Modules may be configured to be addressable storage media and may be configured to play one or more processors. ≪ RTI ID = 0.0 > Thus, by way of example, the terms 'to', 'to', 'to block', 'to module' may refer to components such as software components, object oriented software components, class components and task components Microcode, circuitry, data, databases, data structures, tables, arrays, and the like, as well as components, Variables. The functions provided in the components and in the sections ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ' , '~', '~', '~', '~', And '~' modules with additional components.

The present invention relates to a method for generating an information code capable of defining an object. The existing object recognition method is performed by feature point extraction, descriptor extraction, and matching step. It takes a long time to extract and match the descriptor compared to the feature point extraction time, so it is suitable for providing real-time service in the mobile environment I did. However, according to the present invention, an object is defined as an information code and stored in a database, and when the object is recognized, the recognition speed of the object can be improved by comparing the information code of the object with the information code stored in the database.

1 to 7 are views for explaining an image coding method of an object according to an embodiment of the present invention.

As shown in FIG. 1, an image encoding method of an object according to an embodiment of the present invention includes extracting feature points of an object (S10), selecting an optimal feature point among the extracted feature points (S20) And generating an information code (S30) for defining an object.

Step S10 of extracting feature points of the object can be performed by a feature point extraction algorithm such as the existing SURF algorithm or FAST algorithm.

Referring to FIG. 2, a feature point extraction result can be confirmed from an image of an object. In Fig. 2, the marked portions are minutiae points of the object.

The SURF algorithm or the FAST algorithm extracts all feature points present in the object. In the present invention, step S20 is performed to select an optimal feature point among the extracted feature points in order to increase object recognition efficiency.

The step S20 of selecting the best feature point among the extracted feature points includes a step S22 of classifying the concentrated feature points and a step S24 of selecting feature points closest to the color mean value among the feature points. The color average value means an average of eight points surrounding the feature point and an RGB value of the feature point. It is possible to increase the efficiency of object recognition by recognizing objects by selecting a predetermined number of optimal feature points.

According to the embodiment, four to eight optimal feature points can be selected.

Referring to FIG. 3, the result of selecting the best feature point among the feature points shown in FIG. 2 can be confirmed. One of the five feature points on the upper left is selected as an optimal feature point, one of the three feature points on the lower left side is selected as an optimal feature point, and one feature point on the upper right side is selected as an optimal feature point.

A step S30 of generating an information code for defining an object using the minutiae points includes a step S32 of calculating the distance between the minutiae points, a step S34 of calculating the color average value around the minutiae points, (S36) of obtaining the division information of the feature points, and an information code generation step.

In the step S32 of calculating the distance between the minutiae points, the distance between the minutiae points can be calculated as an absolute distance using the resolution of the image. The distance between the minutiae points is the first unique field of the information code. As described above, when 4 to 8 optimal feature points are selected, the number of combinations of the distance between the feature points is 6 to 28, so matching for object recognition can be performed quickly.

According to the embodiment, the distance between the first feature point and the second feature point may be stored in the database in the form of "1-2: 100 ".

Referring to FIG. 4, the distance between four minutiae points, that is, distance information between six minutiae points is calculated.

In step S34 of calculating the color average value around the minutiae points, the color average value around the minutiae points is calculated using the average of the RGB values of the eight points surrounding the minutiae and the minutiae as shown in Fig. The color mean value around the minutiae is the second unique field of the information code.

According to an embodiment, an information code can be generated that defines an object using the distance between the minutiae and the color mean value around the minutiae. By using the two fields for object recognition, object recognition can be performed in real time in a mobile environment.

In step S36, the resolution of the image may be obtained from the image of the acquired object in the step of acquiring the resolution of the image, the context information of the object, and the division information of the minutiae points.

The context information of the object may include position information of the object, and the position information of the object may be acquired by a position recognition device such as GPS.

The partition information of the feature points includes position information of the feature points in the acquired image of the object, and the position information of the feature points can be obtained by dividing the image of the object by a predetermined method to see how the feature points are divided.

According to the embodiment, the image segmentation of the object for the segmentation information of the minutiae points can be obtained by eight or sixteen segments, and three segmented positions can be used in the order in which the minutiae are distributed in large numbers. This enables faster object recognition by reducing unnecessary seeking operations.

Referring to FIG. 6, the result obtained by dividing the extracted feature point into 8 parts can be confirmed. In FIG. 6, when three divisional positions are selected in the order in which the minutiae are widely distributed, positions 3, 7, and 1 are selected. In an embodiment, the division information of the minutiae points may be stored in N_abc form. N means the number of divisions, and abc means a dividing position where many characteristic points are distributed. Therefore, in the case of FIG. 6, it can be stored in the form of 8_371.

In the information code generation step S38, the information code includes at least one of minutiae calculated in the previous step, distance between minutiae points, color average value around minutiae points, resolution of image, As shown in FIG. The information is exemplary and may include additional information as needed to generate an information code.

Referring to FIG. 7, the information code generated by the image coding method according to the embodiment of the present invention can be confirmed. The x and y coordinates of the four optimal feature points are recorded in the feature point field, the distances between the feature points 1 to 4 are recorded in the distance field between the feature points, and the color average value of four optimal feature points Is recorded. In the resolution field, the resolution of the image is recorded. In the context information field, the position information of the object is recorded. In the division information field, three pieces of division position information having the number of division and the minutiae point are recorded.

Objects in which the information code is generated by the above method are stored in the database and can be used when the user recognizes the object.

8 is a view for explaining an object recognition method according to an embodiment of the present invention.

As shown in FIG. 8, an object recognition method according to an embodiment of the present invention includes extracting feature points of an object (S110), extracting an information code defining an object using feature points (S120) And searching for an object from the database using the code (S130).

Step S110 of extracting feature points of the object may be performed similarly to step (S10) of extracting feature points of the object of Fig.

Step S120 of extracting an information code defining an object using the minutiae points may be performed similarly to step S30 of generating an information code defining an object using the minutiae of Fig. That is, the information code may include at least one of the optimum feature points, the distance between the feature points, the color average value around the feature points, the resolution of the image, the context information about the object, and the division information of the feature points.

Step S130 of searching an object from the database using the extracted information code includes a first filtering step S131 using context information around the object, a second filtering step using the division information of the feature points S132, (S133) comparing the distance between the minutiae points, a step (S134) comparing the distance between the minutiae points, and a step S135 comparing the color average values around the minutiae points.

Among the many object information stored in the database, primary filtering is performed through the context information of the object including the object location information, and secondary filtering is performed through the partition information of the minutiae points, thereby enabling faster searching.

After the filtering is performed, matching is performed using the optimal feature points, the distance between the feature points, and the color average value around the feature points, and the matching rates are calculated. If the average value of the matching rate is higher than the predetermined matching rate, it is determined that the object recognition is successful. If the average value of the matching rate is lower than the predetermined matching rate, the object recognition is determined to be failed. The predetermined matching rate can be set variously as needed.

FIG. 9 shows an object recognition apparatus 100 according to an embodiment of the present invention.

9, the object recognition apparatus 100 according to the embodiment of the present invention includes a feature point extraction unit 10, an information code generation unit 20, a database 30, and an object recognition unit 40 . The image encoding method and object recognition method of the object described above can be performed by the object recognition apparatus 100 shown in FIG.

The feature point extraction unit 10 can extract the feature points of the object.

The information code generating section 20 includes a distance calculating section 21, a color average value calculating section 22, a resolution calculating section 23, a position information recognizing section 24 and a minutia point position analyzing section 25.

The distance calculating unit 21 can calculate the distance between the minutiae points and the color average value calculating unit 22 can calculate the color average value around the minutiae. The resolution calculating unit 23 calculates the resolution of the image of the object, and the positional information recognizing unit 24 can obtain the positional information of the object. According to the embodiment, the location information recognizing unit 24 may include GPS equipment. The minutia point position analyzing unit 25 can divide the image of the object and analyze the split position where the minutiae are widely distributed.

The database 30 may store an information code of an object generated by the information code generating unit 20. [

The object recognition unit 40 can perform object recognition by comparing the information code of the object generated by the information code generation unit 20 with the information code stored in the database.

The image encoding method and the object recognition method of an object according to an embodiment of the present invention can be stored in a computer-readable recording medium that is manufactured as a program to be executed in a computer. The computer-readable recording medium includes all kinds of storage devices in which data that can be read by a computer system is stored. Examples of the computer-readable recording medium include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage, and the like.

It is to be understood that the above-described embodiments are provided to facilitate understanding of the present invention, and do not limit the scope of the present invention, and it is to be understood that various modifications may be made within the scope of the present invention. For example, each component shown in the embodiment of the present invention may be distributed and implemented, and conversely, a plurality of distributed components may be combined. Therefore, the technical protection scope of the present invention should be determined by the technical idea of the claims, and the technical protection scope of the present invention is not limited to the literary description of the claims, The invention of a category.

10: Feature point extraction unit
20: Information code generating section
30: Database
40: Object recognition unit

Claims (19)

Extracting feature points of an object; And
Generating an information code defining the object using the minutiae points;
/ RTI >
The step of generating the information code
Calculating a distance between the minutiae points;
/ RTI >
The method according to claim 1,
After extracting minutiae points of the object
Selecting optimal feature points from the extracted feature points;
The method comprising the steps of:
Extracting feature points of an object;
Selecting optimal feature points from the extracted feature points; And
Generating an information code defining the object using the minutiae points;
/ RTI >
Wherein the step of selecting an optimal feature point among the extracted feature points comprises:
Classifying the minutiae among the extracted minutiae; And
Selecting a feature point closest to the color average value among the classified feature points as the optimal feature point;
/ RTI >
delete The method according to claim 1,
The step of generating the information code
Calculating a color average value around the feature points;
The method comprising the steps of:
6. The method of claim 5,
Generating an information code of the object using a distance between the minutiae points and a color average value around the minutiae points;
The method comprising the steps of:
6. The method of claim 5,
The step of generating the information code
Obtaining at least one of resolution of an image, context information of the object, and division information of the minutiae points;
The method comprising the steps of:
8. The method of claim 7,
Using at least two or more of the extracted feature points of the object, the distance between the feature points, the color average value around the feature points, the resolution of the image, the context information about the object, Generating a code;
The method comprising the steps of:
Extracting feature points of an object;
Extracting an information code defining the object using the minutiae points; And
Searching the object from the database storing the image using the extracted information code;
/ RTI >
The step of generating the information code
Calculating a distance between the minutiae points;
/ RTI >
Extracting feature points of an object;
Extracting an information code defining the object using the minutiae points; And
Searching the object from the database storing the image using the extracted information code;
/ RTI >
The information code
Wherein at least two of the optimal feature points of the object, the distance between the feature points, the color average value around the feature points, the resolution of the image, the context information about the object, and the division information of the feature points are included.
11. The method of claim 10,
The step of searching for the object from the database in which the image is stored using the extracted information code
A first filtering step using context information around the object;
A second filtering step using division information of the minutiae points;
Comparing the optimal feature points;
Comparing distances between the minutiae points; And
Comparing color mean values around the feature points;
/ RTI >
A recording medium on which a program for performing the method according to any one of claims 1 to 3 and 5 to 11 is recorded.
A feature point field for storing feature points of an object;
A feature point distance field storing a distance between the feature points of the object; And
A feature point color average value field for storing a color mean value around the feature points;
The information code of the object to be defined, including.
14. The method of claim 13,
A resolution field for storing the resolution of the image;
A context information field for storing context information about the object; And
A division information field for storing division information of the minutiae points of the object;
The information code of the object that is further defined.
A feature point extracting unit for extracting feature points of the object from an image of an input object; And
An information code generating unit for generating an information code defining the object from the extracted minutiae;
/ RTI >
The information code generation unit
A distance calculating unit for calculating a distance between the extracted minutiae points;
The object recognition apparatus comprising:
A feature point extracting unit for extracting feature points of the object from an image of an input object; And
An information code generating unit for generating an information code defining the object from the extracted minutiae;
/ RTI >
The information code generation unit
A color average value calculation unit for calculating a color average value around the minutiae points;
The object recognition apparatus comprising:
17. The method of claim 16,
The information code generation unit
A resolution calculating unit for calculating a resolution of an image of the object;
A position information recognition unit for acquiring position information of the object; And
A feature point position analyzer for analyzing the position of the feature points in the image of the object;
The object recognition apparatus comprising:
A feature point extracting unit for extracting feature points of the object from an image of an input object;
An information code generating unit for generating an information code defining the object from the extracted minutiae; And
A database storing the generated information codes;
The object recognition apparatus comprising:
19. The method of claim 18,
An object recognition unit for searching an image of the input object in the database;
Further comprising:
Wherein the object recognition unit compares an information code of an image of the input object with an information code stored in the database and searches for the information code.

Images